Pith. sign in

REVIEW 2 major objections 5 minor 14 references

Results from IceCube Searches for High-energy Neutrinos Coincident with Gravitational-Wave Alerts in LVK O4

T0 review · 2 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read IceCube's real-time follow-up of LIGO-Virgo-KAGRA's fourth observing run found no high-energy neutrino emission coincident with any gravitational-wave alert through April 1, 2025, with a population-level $p$-value of 0.15 consistent with…

desk verdict O4 GW-neutrino null result from IceCube is solid and the latency improvements are real; the paper needs upper-limit numbers and a trial-factor caveat, but it deserves review. read the letter →

arxiv 2507.08131 v1 pith:W24NPFMZ submitted 2025-07-10 astro-ph.HE

Zsuzsa Marka , Jessie Thwaites , Justin Vandenbroucke , Doğa Veske (for the IceCube Collaboration) This is my paper · ORCID
classification astro-ph.HE
keywords high-energyneutrinosgravitational-waveastronomymultimessengerreal-timefollow-upcompactbinarymergersIceCubeneutrinoupperlimitsLIGO-Virgo-KAGRA
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper reports that IceCube's real-time follow-up of gravitational-wave alerts from the LIGO-Virgo-KAGRA fourth observing run (O4) found no high-energy neutrino emission coincident with any of the 208 significant and 2,292 low-significance alerts issued through April 1, 2025. The central statistical claim is that the distribution of $p$-values from the Bayesian LLAMA pipeline is consistent with the uniform distribution expected from background, with a population-level $p$-value of 0.15. Individual events did show low $p$-values, including two that triggered electromagnetic follow-up campaigns, but further analysis showed the apparent coincidences were not associated with the gravitational-wave events. The same null result holds for an extended search window of 14 days applied to the five alerts likely to contain a neutron star. These searches set upper limits on the neutrino flux and on the isotropic-equivalent energy emitted in neutrinos.

What carries the argument

The argument is carried by two independent coincidence searches over the Gamma-ray Follow-Up (GFU) track-event stream, which provides low-latency high-energy neutrino candidates. LLAMA computes a Bayesian odds ratio using astrophysical priors (uniform time and sky location, $r^{-2}$ distance prior, log-uniform isotropic energies) and converts it to a frequentist $p$-value against a background built from scrambled neutrinos and the O4 gravitational-wave skymap set. The UML search builds an extended likelihood over the HEALPix pixelized sky with the GW skymap applied as a spatial penalty, and computes $p$-values from background-only pseudo-experiments. The population statement comes from combining per-event $p$-values and comparing the distribution to the uniform expectation, using an updated background that incorporates O4 sky localization information.

What would settle it

A confirmed gravitational-wave alert with a coincident high-energy neutrino whose joint $p$-value falls far below the background expectation, or a full-O4 population test with $p$-value well below 0.01, would refute the no-emission conclusion. A simpler technical check is a signal-injection run that adds fake neutrino events at known positions to the GFU stream and tests whether both pipelines recover them at the claimed false-alarm rate.

Watch

Extended reading notes

Core claim

The central claim is that none of the gravitational-wave alerts from the first part of O4 is accompanied by high-energy neutrino emission distinguishable from background. Two independent real-time searches, LLAMA and the unbinned maximum likelihood (UML) 'generic transient' search, examined each alert in a $\pm 500$ s window around the merger time; for five neutron-star-containing candidates, UML also searched a $[-0.1, +14]$ day window. The observed $p$-value distributions and the population test ($p=0.15$) are consistent with background, and upper limits are placed on the neutrino flux and isotropic-equivalent energy for each alert. The paper also documents the switch to fully automated GCN notices, which reduced the median response latency from 56 minutes to 22 minutes.

Load-bearing premise

The entire null result rests on the scrambled-neutrino background: if scrambling the GFU track sample against the O4 gravitational-wave skymaps does not faithfully reproduce the spatial and temporal correlations of real background, then the quoted $p$-values, including the 0.15 population result, could be biased.

Editorial extensions

If this is right

  • If the null result persists through the remainder of O4, the accumulated population will place tighter upper limits on the neutrino flux and isotropic-equivalent energy emitted by compact binary mergers.
  • The automated GCN notice stream means that electromagnetic observatories now receive neutrino coincidence information within roughly 22 minutes of a merger, a latency short enough to enable prompt follow-up of joint candidates.
  • The consistency of the $p$-value distributions with background constrains the fraction of O4 compact binary mergers that emit detectable high-energy neutrinos, limiting hadronic jet models.
  • The two sub-1% events illustrate the real-time program in action: low-$p$-value alerts are circulated and followed up quickly, even when later skymap updates or optical observations show the coincidence was not a true association.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the null result continues to the end of O4, the combined population test could reach sensitivity to exclude optimistic models where a few percent of mergers produce bright neutrino flares; the paper does not extrapolate this far.
  • The scrambling-based background could be validated independently by injecting simulated neutrino events into the GFU stream and checking that both pipelines recover them at their claimed significance; the paper reports upper limits but does not show such an end-to-end test.
  • The 22-minute latency is dominated by the fixed $\pm 500$ s window and by waiting for the next GFU track; an early-warning or shorter-window variant would cut this further, though with reduced sensitivity.
  • Joint GW+neutrino skymaps, mentioned as ongoing work, would likely reduce the number of misleading low-$p$-value alerts by suppressing chance coincidences between unassociated events.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. This ICRC proceedings paper describes IceCube's real-time follow-up of LIGO-Virgo-KAGRA O4 gravitational-wave alerts using two independent pipelines: LLAMA, a Bayesian analysis with astrophysical priors, and UML, an unbinned maximum-likelihood search. It reports cumulative results through April 1, 2025: 208 significant and 2292 low-significance alerts were processed, an automated GCN Notice stream reduced median response latency from 56 to 22 minutes, and no significant neutrino excess was found. The LLAMA p-value distribution is consistent with a uniform distribution, a population test gives p = 0.15, and the UML p-values agree with the background expectation. The paper also discusses two individual events with low p-values and states that 22 joint candidates below 1% were released for electromagnetic follow-up. The abstract claims that upper limits are set on flux and isotropic-equivalent energy, but no numerical upper limits appear in the body.

Significance. If the results are taken as stated, the paper provides a useful multi-messenger constraint: it extends IceCube's GW follow-up through roughly two years of O4, includes both significant and subthreshold alerts, and documents a practical improvement in real-time follow-up via automated GCN Notices. The central null claim is reasonably supported by the reported p-value distributions and by the population test p-value of 0.15, and the paper is transparent about the fact that individual p-values are not corrected for trial factors. However, the quantitative upper-limit result promised in the abstract is not actually presented, and the population test that underlies the strongest null statement is not specified. Both issues are fixable but need to be addressed before the paper can be considered complete.

major comments (2)
  1. [§5] The sentence reporting a population test p-value of 0.15 is the strongest quantitative null result in the paper, but the test is not defined: the text does not state the test statistic, the null distribution, how multi-event skymap updates are handled, or whether the non-uniform background expectation described for Figure 4 is accounted for. No reference is given for this test. Please specify the test construction or cite a methods paper; as written, this p-value cannot be independently assessed.
  2. [Abstract and §6] The abstract states that the paper sets upper limits on both flux and isotropic-equivalent energy emitted in neutrinos, but no numerical upper limits, upper-limit definitions, or representative values appear anywhere in the text. Section 3 mentions a time-integrated flux sensitivity range in GCN Notices, but the paper itself does not report any of these quantities. This makes the stated deliverable unverifiable. Please add a table of representative limits (or an aggregate limit) and define the assumed spectrum, time window, and distance/energy conventions, or explicitly state that numerical limits are only available in the archived GCN Notices and do not form part of this paper.
minor comments (5)
  1. [§2] The criteria HasNS > 0.5 and HasRemnant > 0.5 are used to select events for the extended time window, but these LVK parameters are never defined or referenced. Please add a brief definition or a citation to the relevant LVK alert documentation.
  2. [§5] The discussion of Figure 4 explains why the background p-value expectation is non-uniform, but it would help to state explicitly how the observed black points in the right panel are combined into a global test statistic, or to state that no global test statistic is used.
  3. [§4] For the S230904n example, the text says LLAMA measured a p-value of 0.0037 and that the coincident neutrino was followed up by ZTF, but it does not report the released neutrino direction or angular uncertainty. A brief coordinate or a pointer to the relevant GCN Circular would make the example self-contained.
  4. [Throughout] The text contains several missing spaces and typographical issues, e.g., 'throughobservation', 'Highenergy', 'coincidingwithobservation', and 'counterpartswerealso'. A copyediting pass is recommended.
  5. [§6] The statement that 22 joint candidates had p-values below 1% is interesting, but no list, table, or distribution of these candidates is given. Even a small table with event IDs, pipeline, p-value, and release status would improve traceability without changing the null conclusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the O4 null result is supported by background-scrambled p-value comparisons, not by any parameter fitted to the neutrino data or by a self-citation chain.

full rationale

The paper's central claim is an empirical null result: no significant high-energy neutrino emission is found coincident with LVK O4 gravitational-wave alerts. The p-values for both LLAMA and UML are computed by comparing observed test statistics to background distributions built from scrambled neutrinos and the gravitational-wave skymap set (Section 2), and the population statement (p=0.15, Section 5) is a standard stacked comparison of the observed p-value distribution to the expected uniform background. No free parameter is fit to the O4 neutrino data: the E^-2 spectrum, the ±500 s window, and the [−0.1,+14] day extended window are fixed modeling assumptions stated in the text, not outputs of the analysis. The citations to previous LLAMA and UML papers are method references; the validity of the null conclusion does not reduce to an unverified self-citation, because the inference is calibrated against background pseudo-experiments and the reported result is a non-detection. Limitations such as the absence of trial-factor corrections for the 22 sub-1% candidates and the missing numerical upper limits are presentation or completeness concerns, not circularity. No equation or construction in the paper defines its output in terms of its input, and no fitted quantity is relabeled as a prediction.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new entities. The central null result rests on standard search methodology and a set of modeling choices (time windows, spectral index, prior shapes) that are hand-picked from the literature but not fitted to the O4 neutrino data. These choices affect the sensitivity and the interpretation of upper limits, but the absence of a significant signal is robust to moderate changes in the assumptions.

free parameters (4)
  • Coincidence search time window = ±500 s (1000 s total)
    Hand-chosen based on GRB emission studies (Ref. [7]); determines which neutrinos count as coincident. Not fitted to O4 data.
  • Extended emission search window = [-0.1, +14] days
    Hand-chosen for NS-containing events based on long-lived remnant models (Ref. [3]).
  • Neutron-star membership thresholds = >50% for p_BNS+p_NSBH, HasNS, HasRemnant
    Hand-chosen criteria to select events for the extended search; affects 5 analyzed events.
  • Assumed neutrino spectrum for upper limits = E^-2
    Standard assumption for IceCube point-source searches; affects computed flux upper limits, not fitted to the search data.
assumptions (6)
  • domain assumption The ±500 s search window, motivated by GRB studies [7], encompasses the neutrino emission timescale for compact binary mergers.
    If neutrino emission occurs outside this window (except the 14-day extended search), the analyses will miss it. The triangular time profile is also an assumed shape.
  • domain assumption Scrambled GFU data and the GW skymap set provide an unbiased background model for computing p-values.
    Both pipelines derive significance from background-only pseudo-experiments based on scrambled neutrinos; any mismatch with the true null distribution would bias the quoted p-values.
  • domain assumption The GFU track event sample and its angular/energy reconstruction are appropriate and calibrated for this search.
    The search uses only high-energy track-like events from the GFU stream (rate 6-7 mHz) with stated localization of order 1 square degree; the null result is conditioned on this sample.
  • domain assumption The assumed E^-2 neutrino spectrum is used for upper limits and sensitivity.
    Upper limits on flux and isotropic energy are computed under E^-2; a harder or softer spectrum would change the limits, but the null detection result is independent.
  • domain assumption LLAMA's astrophysical priors (uniform in time, uniform sky location, r^2 distance prior, log-uniform emitted energy) are reasonable descriptions of the source population.
    These priors enter the Bayesian odds ratio; if they are strongly wrong, the odds ratio significance could be miscalibrated, though the frequentist p-value from scrambles partially guards against this.
  • domain assumption LVK's alert stream and p_astro values provide a faithful catalogue of candidate compact binary mergers.
    The analysis responds to LVK alerts and uses p_astro as an input; the sample is defined by LVK's release threshold, which the paper notes is 2/day per pipeline.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Results from IceCube Searches for High-energy Neutrinos Coincident with Gravitational-Wave Alerts in LVK O4." pith.science (2026). https://pith.science/paper/W24NPFMZ

@misc{pith2026250708131,
  author       = {Pith},
  title        = {Pith review of: Results from IceCube Searches for High-energy Neutrinos Coincident with Gravitational-Wave Alerts in LVK O4},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W24NPFMZ}},
  note         = {Machine review of arXiv:2507.08131}
}
read the original abstract

Mergers of compact objects, binary black holes and mergers including at least one neutron star, are a predicted source of high-energy neutrinos. These astrophysical events are now routinely detected through observation of their gravitational wave signature and, at least in one instance, their electromagnetic counterparts were also detected. Particles accelerated during the coalescence of compact objects may also interact to produce high-energy neutrinos, which have yet to be detected, but observations are ongoing. The LIGO-Virgo-KAGRA Collaboration publicly releases information on candidate gravitational wave events from compact binary coalescences in low latency during the current observing run (O4). To aid the electromagnetic follow-up, using data from the IceCube Neutrino Observatory, we search, in real time, for neutrinos spatially and temporally coincident with these gravitational wave candidate events using a time window of 1000 seconds centered on the merger time. We use two methods, both of which have been previously used to search for neutrino emission from gravitational-wave transients: an unbinned maximum likelihood analysis applied to significant alerts and a Bayesian analysis with astrophysical priors, applied to both significant and low-significance alerts. In addition, we search for long-duration neutrino emission up to 14 days after the merging of binaries containing a neutron star. We report analysis results determined in real time for these searches, and set upper limits on both flux and isotropic-equivalent energy emitted in neutrinos.

Figures

Figures reproduced from arXiv: 2507.08131 by the authors.

Figure 1
Figure 1. Latency distribution for the joint UML and LLAMA GW follow-up for the high significance LVK alerts. In O3, all events were sent as GCN Circulars, while in O4 most events have been sent automatically via GCN Notice. The top panel shows the IceCube-only latency, while the bottom panel shows the total latency (with reference to the merger time). Only the first GCN for each Significant event in O4 is shown here, and thi… view at source ↗
Figure 2
Figure 2. Illustration of the evolution of the S250206dm (55% NSBH/37% BNS/8% Terr.) skymap in real time. With the 3-Initial skymap, UML found a 𝑝-value of 0.008 and LLAMA 0.011. For LLAMA, neutrinos 4 and 8 reached the threshold of 10% for release. The most significant 𝑝-value (0.008) from UML corresponded to neutrino 8, which prompted the team to release a GCN Circular. At the time of the GW event, Virgo was being brought o… view at source ↗
Figure 3
Figure 3. 𝑝-value distribution of the coincidences analyzed by the LLAMA pipeline in O4 until the April 2025 break. The orange band shows the expected uniform distribution and the standard deviation per bin [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Background TS and p-value distributions for the unbinned maximum likelihood search for GW events in O1-O4 through April 1, 2025. (Left) The background-only TS distribution for each GW event (colored histograms). The area of the skymap (color bar) affects how many backg…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

  1. [1]

    Abbott,et al., Physical Review X13no

    LIGO, Virgo, and KAGRA Collaborations, R. Abbott,et al., Physical Review X13no. 4, (Oct., 2023) 041039

    work page 2023

  2. [2]

    S. S. Kimura, K. Murase, I. Bartos, K. Ioka, I. S. Heng, and P. Mészáros,Phys. Rev. D98 no. 4, (Aug., 2018) 043020

    work page 2018

  3. [3]

    Fang and B

    K. Fang and B. D. Metzger,Astrophys. J.849 no. 2, (Nov., 2017) 153

    work page 2017

  4. [4]

    Abbasiet al., ApJ 944 no

    IceCubeCollaboration, R. Abbasiet al., ApJ 944 no. 1, (Feb., 2023) 80

    work page 2023

  5. [5]

    IceCubeCollaboration, M. G. Aartsenet al.,ApJL 898 no. 1, (July, 2020) L10

    work page 2020

  6. [6]

    Abbasiet al., ApJ 910 no

    IceCubeCollaboration, R. Abbasiet al., ApJ 910 no. 1, (Mar., 2021) 4

    work page 2021

  7. [7]

    Baretet al.,Astropart

    B. Baretet al.,Astropart. Phys.35no. 1, (Aug., 2011) 1–7

    work page 2011

  8. [8]

    Bartos, D

    I. Bartos, D. Veske, A. Keivani, Z. Márka, S. Countryman, E. Blaufuss, C. Finley, and S. Márka,Phys. Rev. D100 (Oct, 2019) 083017

    work page 2019

Show all 14 references
  1. [9]

    Countryman, A

    S. Countryman, A. Keivani, I. Bartos, Z. Marka, T. Kintscher, R. Corley, E. Blaufuss, C. Finley, and S. Marka,arXiv e-prints(Jan., 2019) arXiv:1901.05486

    2019 arXiv

  2. [10]

    Countryman,Computational Methods in Multi-Messenger Astrophysics using Gravitational Waves and High Energy Neutrinos

    S. Countryman,Computational Methods in Multi-Messenger Astrophysics using Gravitational Waves and High Energy Neutrinos. PhD thesis, Columbia University, 2023

    2023

  3. [11]

    Braun, J

    J. Braun, J. Dumm, F. De Palma, C. Finley, A. Karle, and T. Montaruli,Astropart. Phys.29 (2008) 299–305

    2008

  4. [12]

    Braun, M

    J. Braun, M. Baker, J. Dumm, C. Finley, A. Karle, and T. Montaruli,Astropart. Phys.33 (2010) 175–181

    2010

  5. [13]

    K. M. Górskiet al., The Astrophysical Journal622no. 2, (Apr, 2005) 759

    2005

  6. [14]

    Abbasiet al., Astrophys

    IceCubeCollaboration, R. Abbasiet al., Astrophys. J.910 no. 1, (2021) 4. 8 IceCube GW Follow-up in LVK O4 Full Author List: IceCube Collaboration R. Abbasi16, M. Ackermann63, J. Adams17, S. K. Agarwalla39, a, J. A. Aguilar10, M. Ahlers21, J.M. Alameddine22, S. Ali35, N. M. Ami...

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.